Dislocation arrangement in copper single crystals as a function of strain

Abstract

Deformed copper single crystals were sectioned both parallel and perpendicular to the primary slip plane by spark erosion and then electropolished to obtain thin foils. Transmission electron microscopy was used to determine both the large scale (0.1 mm) features of the dislocation array and also details of characteristic configurations. Comparison of the observations from the mutually orthogonal sections made it possible to construct a three-dimensional model of the internal structure. An extensive Burgers vector survey was conducted by a bright field technique making use of dislocation disappearance criteria. Observations were almost entirely restricted to specimens with two seeded orientations, one for single and the other for double slip. Crystals were examined in each of the three stages of the work hardening curve. Dislocation loss during thinning is not considered to be of major importance and reasons for this opinion are given. In stage I primary screw dislocations were absent from the foils and it is concluded that when screws of opposite sign pass on sufficiently close slip planes ($\lesssim$200 $\overset{\circ}{\mathrm A}$) they can cross-slip and annihilate. Under similar conditions edge dislocations form stable arrays and these accumulate more edges through glide polygonization and thus form extensive walls perpendicular to the primary plane. A very high density of secondary dislocations was observed throughout stage II and this can be reconciled with the macroscopic secondary strain because the secondary slip distance was found to be very short. It is proposed in this paper that secondary slip occurs in the regions of high internal stress at the heads of pile-ups and tests of this proposal are applied. No classical pile-ups were observed; instead the secondary slip converts the pile-up into a complex tangle which is stable on unloading. Some of the predictions of a work hardening theory arising out of these observations (Hirsch 1964) are compared with experiment. Lomer-Cottrell dislocations were found to be unstable at room temperature. Faint lines of contrast along close packed directions in the primary plane were shown to be elongated faulted loops. Their observed maximum width gives, in principle, an estimate for the stacking fault energy of copper.